On a mathematical model of immune competition